-------------------------------------------------------------------------------------------------------------------------------

2 시스템 사고1  - 단선적 시각과 시스템사고

One Way Linear Causation Thinking vs. System Thinking

-------------------------------------------------------------------------------------------------------------------------------

 

1. 단선적 사고의 문제점 Problem of one-way linear causation thinking

 

정책실패(policy failure) 정책으로 해결하고자 했던 문제를 해결하지 못하거나, 목표를 달성하였다 하더라도 정책이 의도하지 않았던 부작용이 정책으로 인한 이익보다 사회전체적으로 손실을 가져오는 경우를 말한다.

 

정책실패의 :

빈곤가난의 퇴치 - 성장위주의 경제정책 : 부의 불균형, 지역간 격차, 환경문제

부동산투기 - 토초세 : 무차별토지이용의 촉진, 도시공간의 효율적 개발 봉쇄

주택문제 - 신도시 건설: 수도권 문제

농산물가격상승 - 농산물 수입정책 : 농산물 가격 폭등 폭락의 악순환

쓰레기문제 - 종량제: 대기오염

 

이와 유사한 정책실패의 예를 한가지씩 들고 설명해 보시오

 

 

정책실패의 원인(김도훈 문태훈 김동환, 1999):

 

부분간의 상충 trade off among sectors

시간간의 상충 short term and long term consequences

 

비체계적접근 non systematic approach– 문제의 즉흥적 해결. 대증적 해결

단기적시각 short term perspective- 장기적 관점의 결여

비통합적 non integrative, neglecting complexity of causal relationship - 부분적 시각 복잡성의 간과

 

단선적 사고 : 사회문제의 상호연관성의 증대

 정태적 : 단일 방향의 인과관계

 단기적 :

 부분적

 

시스템사고 수필 모음

Donella H. Meadows, Sustainability Institute

http://www.sustainer.org/

http://www.pcdf.org/meadows/default.htm

http://www.sustainer.org/dhm_archive/

http://iisd1.iisd.ca/pcdf/meadows/default.htm

 

Thinking Capacity

 

 

content = fact about things. Assimilated largely through memorization

understanding = inventory you hold in your head of relationship between things

                               connections you made, relationship between things

                               mental model – ex. Hot things burn you.

Thinking capacity = ability to

1)       construct mental model

2)       make comparison b/w models

3)       resolve conflict b/w these models into appropriate adjustments to currently held mental models

 

construction, comparison, resolution

 

 

A Simple Independent System

Mature tree behavior responding to STEP increase in harvesting

 

 



 

System (Andersen, 1997:1-9)

 

1) meaning:

A system is a group of interacting, interrelated, or interdependent components that form a complex and unified whole.

 

A System's components can be

1) physical objects that you can touch: car parts

2) intangible : process, relationships company policies, information flows....

 

people+equipment+process=R&D group < production < sales < company < industry as a whole < economy

 

2) characteristics of systems

(1) Systems parts must all be presents for the system to carry out its purpose optimally

ex) If you can add or subtract a components to a collection without affecting its functioning and relationships, it's just a collection not a system.

Thus, a bole of mixed nuts is not a system but a collection

 

(2) A system's parts must be arranged in a specific way for the system to carry out its purpose

ex) In a bowl of fruit, the orange can go at the bottom, in the middle, or on the top without changing the essential nature of collection of fruit.

However, in a system such as a company, you cannot shift its part around randomly without affecting system's performance

 

(3) System have specific purpose within larger systems.

All systems have a specific purpose in relationship to the larger system in which they are embedded in.

ex) R&D Department. Human circulatory system.

 

(4)System maintain their stability through fluctuations and adjustments.

Left to themselves, systems seek to maintain their stability.

ex) A company does its best to maintain a designated profit margin

Human bodies work to maintain a temperature of about 36.5 degree.

 

System achieve this stability through the interactions, feedback, and adjustments that continually circulate among the system parts, and between the system and its environment.

 

(5) Systems have feedback

Feedback is transmission and return of information.

ex) driving car.

 

A system has feedback within itself. But because all systems are part of larger systems, a system also has feedback between itself and external systems.

 

In some systems, the feedback and adjustment processes happen very quickly, but it may take a long time before the feedback is returned. (ex: sunbath)

 

Feedback is not necessarily transmitted and returned through the same system component - or even through the same system. It may travel through several intervening components within the system first, or return from an external system, before finally arriving again at the component where it started.

 

 

layoff workers may improve financial condition in the short term, but in the long term, it often worsen financial condition which often lead to further layoff

 

3. System Thinking

 

We tend to focus on linear causal relationships rather than circular or mutually causative ones.

Yet, many of the most vexing problems confronting managers and corporations today are caused by a web of interconnected, circular relationships.

 

1) Principles of system thinking

 

In general, system thinking is characterized by following principles(Andersen, 18)

thinking of the big picture

thinking of the big picture

balancing short-term and long term perspectives

recognizing the dynamic, complex, and interdependent nature of systems

taking into account both measurable and nonmeasurable factors

remembering that we are all part of the systems in which we function, and that we each influence those systems even as we are being influenced by them

 

thinking of the big picture

One of the disciplines of systems thinking involves being able to step back from that immediate focus and look at the bigger picture.

To discover the source of a problem, you have to widen your focus to include that bigger system.

 

Maintaining a binocular perspective

Keeping an eye tuned to the biggest relevant picture while the other eye discends into the fray

A good basketball player = tactical skill + court sense; feeling for the whole

Seeing both the forest and the trees

 

 

balancing short-term and long term perspectives

behavior that leads to short-term success or that is prompted by short-term assessment can actually hurt long term success

 

recognizing the dynamic, complex, and interdependent nature of systems

Things change all the time, life is messy, and everything is connected

We need to be aware of all the system's relationships-both within it and external to it.

 

taking into account both measurable and nonmeasurable factors

ex)sales figures and costs vs. morale and customer attitudes

 

We are Part of the System:

Remembering that we are all part of the systems in which we function, and that we each influence those systems even as we are being influenced by them

 

Unintended consequences: problem plaguing us today is an unintended consequence of a solution we implemented yesterday

ex) To reduce cost, one bank shut down a number of tellers on Thursday evening and Saturday morning. But other bank get all the customer who rely on having access to the bank during evening and weekends.

Ex) ammonia refrigerating -> CFC chlorofluorocarbon -> stratosphere ozone destruction

 

Assumptions: sometimes our assumptions are what get us into trouble

 

Value and belief: deeply held values and beliefs can lock us into counterproductive ways of making decisions.

 

In summary,

System thinking emphasizes looking at wholes rather than parts, and stress the role of interconnections. It recognizes that we are part of the system in which we function. It is circular rather than linear language. It focus on closed interdependencies where x influences y, y influences z, and z comes back around to influence x. It offers visual tools, such as causal loop diagram. (Andersen, 20-21)

 

2) Event, Patterns, Structure

 

Structure is overall way in which the system components are interrelated - the organization of a system. Because structure is defined by the interrelationships of a system's parts, and not the parts themselves, structure is invisible.

Understanding system's structure is important because it's system structure that gives rise to all the events and trends that we see happenings in the world around us.

 

(1)Events

We live in an event-focused society

But focusing on events is like wearing blinders. You can only react to each new event rather than anticipate and shape them. What's more, solutions designed at the event level tend to be short lived.

Most important they do nothing to alter the fundamental structure that caused that event.

ex) building burning - fire fighter put out fire

 

By uncovering the elusive systemic structure that drives events, you can begin identifying higher-leverage actions.

 

(2)Patterns

Patterns are trends, or changes in events over time.

Whenever you see a pattern of events, you are getting one step closer to grasping the systemic structure driving that pattern.

ex) sales have been declining over the past few years, several senior engineers have left the company recently...

 

(3)Structure

Thinking at the structure level means thinking in terms of causal connections. It is the structural level that holds the key to lasting, high-leverage change.

 

senior engineers exodus.

 

 

Level of Understanding

 

Action Mode

Time Orientation

Way of Perceiving

Questions you would ask

Events

React

Present

Witness event

What's the fastest way to react to this event now?

Patterns

Adapt

 

Measure or track patterns of events

What kinds of trends or patterns of events seem to ve crcurring?

Structure

Create

Change

Future

Causal loop diagrams and other systems thinking tools

What structures are in place that are causing these patterns?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Anderson, Virginia and Lauren Johnson. 1997. System Thinking Basics. Pegasus Communications, Inc.

 

 

4. 시스템다이내믹스 System Dynamics

 

시스템다이내믹스(System Dynamics, 체계동태학) 동태적이고 순환적 인과관계의 시각으로 (dynamic feedback perspective) 현상을 이해하고 설명하며, 이러한 이해에 기초한 컴퓨터 모델을 구축하여 복잡한 인과관계로 구성된 현상이 어떻게 동태적으로 변해 나가는지를 컴퓨터상에서 실험해 보는 방법론이자 현상을 바라보는 시각이며 준거틀(framework) 이다.

 

System Dynamics is a perspective and framework of understanding reality with dynamic and circular feedback causation perspective and computer simulation methodology to build model with such dynamic perspective of reality and to test dynamic behavior of interest variables over time.

참고: 문태훈. 2002. 시스템다이내믹스의 발전과 방법론적 위상. 시스템다이내믹스연구 3권 2호. .pdf

 

순환적인과관계(Circular Causality), 피드백루프(feedback loop)  피드백루프에 의한 순환적 과정(circular process as feedback loops)

 

1) Dynamic thinking

연구하고자 하는 특정 변수가 시간의 변화에 따라 어떻게 동태적으로 변화해 나가는가에 기본적인 관심을 둔다. , 산업체 고용인력의 증감, 도시의 번영과 쇠퇴, 의료보험 비용의 급격한 상승 연구 대상 변수의 시간에 따른 동태적인 변화가 어떻게 일어나고 있으며, 앞으로는 어떻게 변화해 것인가에 대해 관심을 둔다는 것이다. 따라서, 시스템다이내믹스는 일회적인 사건이나 모델 파라미터의 정확한 측정이나 변수의 추정 값을 구하기보다는 관심의 대상이 되는 변수가 시간의 흐름에 따라 어떤 동태적인 변화의 경향 --안정적, 불안정적 경향, 상하 주기적인 변동, 성장, 쇠퇴, 평형상태의 유지 -- 보이는지에 보다 관심을 둔다(Meadows,1980:31-36).

Dynamic thinking – seeing patterns of behavior over time rather than focusing on events

 

 

2) Endogenous point of view

모든 현상을 내부순환적 환류체계(closed loop thinking, circular feedback system) 관점에서 이해한다는 것이다(김도훈 문태훈 김동환, 1999:36, High Performance, 1994:25-26). , 어떤 변수의 동태적인 변화를 시스템내부에 존재하는 변수들과의 원형의 역동적인 상호작용(circular causation, two-way causation or feedback) 의하여 일어나는 것으로 파악한다. 이때 시스템은 일반적 체제이론에서 말하는 시스템과는 다른 개념을 가진다. 여기서 시스템은 연구대상의 변수가 동태적 변화를 일으키는데 관련되는 변수들의 집합체를 의미한다. 이러한 체제에서는 기존의 개방체제와 폐쇄체제의 구분은 이상 의미를 갖지 못한다. 따라서 변수의 동태적 변화는 시스템 내부의 원인에 의하여 발생하는 것이지 시스템 밖의 외부변수에 의해서 발생하는 것이 아니게 된다. 이런 의미에서 시스템다이내믹스는 내부순환적, 또는 내부지향적 관점(endogenous point of view) 가진다.

 

System as cause thinking – system as cause perspecitve = endogenous point of view

Construct a boundary around a set of elements and interrelationships such that the cause of dynamics exhibited by the system resides within the boundary

 

Ex. Predator-pray oscillation

Fluctuation in the respective population are being caused by many factors – sunspots, water supply, human encroachment, other outside forces

A system as a cause perspective choose to see whether the oscillation could be generated only out of interplay of the predators with prey.

Conscious choice of the investigator to view it as a possibility, and then to explore this possibility – for principle of parsimony

 

Ex. Guest with young child. Child broke vase

Who is responsible – guest vs host. They blame each other

System as cause perspective – view themselves responsible. Each has a complete control over the visiting venue as well as the circumstances of the visit.

 

3) Circular feedback perspective

(closed loop thinking – what’s inside generates the dynamics)

시스템사고에서는 관련된 변수들이 일방향의 단선적인 영향을 주는 것이 아니라 원형의 인과관계에 의하여 동태적인 상호작용을 하고 있는 것으로 파악하기 때문에 원형의 피드백 관점을 가진다(Richardson, 1982:1-2; Meadows, 1980:30-31). [환류체계의 관점에 대한 논의는 Richardson(1991) 참조]

 

아래 [그림1] 일방향의 단선적 인과관계는 학업성취도에 여러 원인들이 일방향으로, 그리고 일회적인 영향을 미치는 구도로 이해되고 있다. 그러나 [그림2] 원형의 피드백관점에서는 학업성취도에 영향을 미치는 변수들이 순환적으로 영향을 주고받는 상호작용 관계에 지속적으로 놓여있음을 나타내고 있다. 종속변수와 독립변수의 구분은 이상 의미가 없게 된다.

 

[그림1] 일방향의 단선적 인과관계

 

 

 

[그림2] 원형의 내부순환적 환류체계(closed loop thinking, circular feedback system)

 

 

순환적인 인과관계에 의한 동태적인 현상들은 주변에서도 많이 찾아볼 있다. 예를 들면 부족한 도로사정으로 도시의 교통혼잡이 야기될 단순한 도로의 확장은 일시적으로 교통혼잡을 완화할 있을지는 몰라도 나아진 도로사정이 결국 많은 자동차를 유인하여 다시 교통체증을 유발하게 된다던지, 농업의 경쟁력을 제고시키기 위하여 정부가 농산물 보조금을 지불할 단기적으로는 경쟁력을 유지시킬 있을지는 모르나 장기적으로는 정부의 재정지원이 농촌의 경쟁력 제고 노력을 둔화시켜 결국 농업의 경쟁력을 더욱 약화시키는 결과를 가져오게 되는 것이다.

 

[그림] 도로와 교통혼잡   road and traffic congestion, more road more congestion              

 

 

[그림] 농촌보조금과 농가구소득 

agricultural subsidies and farmers’ income, more subsidy, less income?

 

 

 

closed loop thinking – what’s inside generates the dynamics

 

Laundary list perspective or laundary list causality

People prioritize or assign weight to the list

Mental models with weighted laundary list structure are static in nature

Causality runs only one-way and the preeminence of causal factor is fixed rather than varying over time

But in reality, virtually every phenomenon is dynamic in nature

 

In a closed loop model, causal relationships are reciprocal. No absolute distinction is maintained between cause and effect.

Each factor is at once cause and effect.

No fixed weightings are assigned

 Teaching quality changes slowly while motivation changes everyday

 

4) Operational thinking

시스템다이내믹스는 사실적 사고(operational thinking) 초점을 둔다. 사실적 사고란 변화가 실제적으로 어떻게 해서 일어나고 있는가 what’s causing change? 초점을 맞추는 사고이다. 시스템 작동의 메카니즘을 파악하고자 하는 사고이다(김도훈 문태훈 김동환, 1999:38; High Performance, 1994:30-31). 미국의  저명한 경제학  논문지에  우유생산을  예측하는  이코노메트릭  모델이  발표된  적이  있다.    모델의 회귀식은 [우유생산 = a1 × (국민총생산)+ a2 × (이자율) + ---- ] 형태를 띠고 있었다. 우유생산은  국민총생산,  이자율과  같은  거시경제적  변수의  함수라는  것이다. 그리고    거시경제적  변수가  우유생산에  영향을  주는  상대적  중요성은  ai  표시하고  있다.  모델은  분명히  우유가  실제적으로  어떻게  생산되는지  나타내  주고  있지  않다.  모델  어디에도  우유생산에  필수적인  젖소()  표현되어  있지  않다.  만일  우리가  실제적  사고를  한다면  젖소()야말로  모델에서  제일  먼저  고려해야   변수일  것이다.  다음으로  젖소  마리가  매년  만들어내는  우유의  평균생산량  중요변수가    것이다.  적어도    변수가  포함되어  있어야  우유생산에  대해  실제적인  언급을    있을  것이다.  사실적  사고란  추상적이거나  수학적인  모델을  가지고  우유생산을  예측하려는  사고가  아니다.  사실적 사고란  실제  우유생산 과정에  무엇이  어떻게  일어나고  있는가를  파악하려는  사고이다.  

 

Operational thinking

Milk production = a1*GNP + a2*Interest rate…

Does this equation help us to find ways to increase milk production?

The equation does not purport to represent how milk is actually produced

Ex. Number of cows, milk per cow – cow’s productivity.

Operational thinking helps you to identify the levers for change that actually are available to the change agent.

To increase milk productioin – bring more cows, longer average duration of employment

 

Dead buffalo model – correlational, not causal model

This does not help to solve problems at all

 

 

High Performance System. 1994. An Introduction to System Thinking.

 

Four system thinking skills – how the world works? World view assumption.

 

1. System as cause thinking – system as cause perspecitve = endogenous point of view

Construct a boundary around a set of elements and interrelationships such that the cause of dynamics exhibited by the system resides within the boundary

 

Ex. Predator-pray oscillation

Fluctation in the respective population are being caused by many factors – sunspots, water supply, human encroachment, other outside forces

A system as a cause perspecitve choose to see whether the oscillation could be generated only out of interplay of the predators with prey.

Conscious choice of the investigator to view it as a possibility, and then to explore this possibility – for principle of parsimony

 

2. closed loop thinking – what’s inside generates the dynamics

Laundary list perspective or laundary list causality

People prioritize or assign weight to the list

Mental models with weighted laundary list structure are static in nature

Causality runs only one-way and the preeminence of causal factor is fixed rather than varying over time

But in reality, virtually every pehnomenon is dynamic in nature

 

In a closed loop model, causal relationships are reciprocal. No absolute distinction is maintained between cause and effect.

Each factor is at once cause and effect.

No fixed weighting are assigned

 Teaching quality changes slowly while motivation changes everyday

 

3. Operational thinking

Milk production = a1*GNP + a2*Interest rate…

Does this equation help us to find ways to increase milk production?

The equation does not purport to represent how milk is actually produced

Ex. Number of cows, milk per cow – cow’s productivity.

Operational thinking helps you to identify the levers for change that actually are available to the change agent.

To increase milk productioin – bring more cows, longer average duration of employment

 

Dead buffalo model – correlational, not causal

 

4. Dynamic thinking – seeing patterns of behavior over time rather than focusing on events